[Pw_forum] Graphene Nanoribbon (Negative phonon frequencies and structure relaxation)
Ari P Seitsonen
Ari.P.Seitsonen at iki.fi
Sat Jan 3 15:12:03 CET 2015
Dear Prabhakar,
Ehmm, I might be wrong, but isn't this a zig-zag ribbon that your input
produces (vertical lines for unit cell; the vertical-to-horisontal ratio
not correct)::
|
| |
C C C C C C C <-- edge
c c c c c c c
... c c c c c c c ...
C C C C C C C <-- edge
| |
The arm-chair would be
| |
|CC |CC CC <-- edge
... cc cc cc cc ...
|CC |CC CC <-- edge
|
Then two issues: The zig-zag ribbon is spin-polarised
(anti-ferromagnetic, the opposite edges with opposite spins) and now that
I start to remember it, it might be that indeed this ribbon turns into two
linear chains (five+ years since those calculations by Tobias...); please
try with a wider one (or indeed the arm-chair ribbon).
Greetings,
apsi
PS Please remember to attach your affiliation to your mails
-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
Ecole Normale Supérieure (ENS), Département de Chimie, Paris
Mobile (F) : +33 789 37 24 25 (CH) : +41 79 71 90 935
On Fri, 2 Jan 2015, Prabhakar Marepalli wrote:
> Dear Dr. Seitsonen,
>
> Thanks for checking my script. You are right that the structure I posted runs in x-direction, but the problems I mentioned are exactly the same for this script too (I used y-notation as it is
> easier to draw and explain the unitcell that way). Also, thanks for pointing about the unitcell plane distance in z-direction. I've done the following changes:
>
> 1) I've increased the lattice constant in z-direction to a very large value of 10 Angstroms, but it didn't change my answers by a lot.
>
> 2) My structure is periodic in all directions, but I've picked the lattice vectors in y,z directions to be very long that the structure essentially acts like a 1D.
>
> 3) I've added the 'ecutrho' parameter and picked a value of ecutrho=12*ecutwfc. This hasn't changed my answers either.
>
> For relaxing, I've made the following changes to my original script:
>
> &control
> calculation='relax'
>
> &ions
> ion_dynamics='bfgs'
>
> &cell
> cell_dynamics='bfgs'
> press_conv_thr=0.4,
> cell_factor=3.0
>
> The above relaxation almost changed my nanoribbon lattice to a square lattice. Can you please suggest if I'm using any of those parameters incorrectly?
>
> Thanks,
> Prabhakar Marepalli
>
>
>
> On Fri, Jan 2, 2015 at 1:16 PM, Ari P Seitsonen <Ari.P.Seitsonen at iki.fi> wrote:
>
> Dear Prabhakar,
>
> Did you check your structure? To me it looks very weird in 'XCrysDen', the planes of the periodic GNR's along z direction only 2 Angstroms apart. And yes, you should relax the
> structure before calculating phonons. And your system is periodic in all directions, and the GNR seems to run along x, not y like I understood from your description below. Please also
> read the instructions about 'ecutrho' when using ultra-soft pseudo potentials or the PAW method.
>
> Good Luck, :)
>
> apsi
>
> -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
> Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
> Ecole Normale Supérieure (ENS), Département de Chimie, Paris
> Mobile (F) : +33 789 37 24 25 (CH) : +41 79 71 90 935
>
>
> On Fri, 2 Jan 2015, Prabhakar Marepalli wrote:
>
> Hello everyone,
>
> I'm trying to compute phonon dispersion properties of graphene nanoribbon
> using Quantum Espresso. My unit cell is a simple armchair structure with 4
> atoms per unit cell. It looks like below:
>
> 0----0
> / \
> 0 0
>
> The above structure is replicated in the y-direction (the periodic
> direction) while the x-direction is confined. Using the above structure,
> here is the outline of the problems I'm running into:
>
> 1) Negative phonon frequencies (not only at gamma point, but at all other
> k-points)
>
> 2) I tried using 'asr' but that doesn't seem to fix the problem.
>
> 3) I understand that the problem might be due to instability of the
> structure (I did notice that the forces on atoms in scf calculation is
> non-zero , forces of the magnitude 0.2 to 0.4).
>
> 4) When I try to relax the structure using 'ion_dynamics', the forces did go
> down but the structure is relaxed so much that it is no longer graphene
> nanoribbon.
>
> Can you please give me some pointers on how to approach this problem, i.e.
> how do I get rid of negative frequencies while relaxing the structure that
> constrains it to look like a nanoribbon ?
>
> Thanks for your time!
>
> Here are my input files:
>
> pw.x (this doesn't include the ion_dynamics relaxation).
>
> &control
> calculation='scf'
> restart_mode='from_scratch',
> tstress = .true.
> tprnfor = .true.
> prefix='gnr_arm',
> pseudo_dir = '/work/02420/pm24229/QE/QE/pseudo/',
> outdir = '/home1/02420/pm24229/tmp/',
> /
> &system
> ibrav= 0, nat= 4, ntyp= 1, nbnd= 20,
> ecutwfc =25.0, occupations='smearing', smearing='marzari-vanderbilt',
> degauss=0.02
> /
> &electrons
> diagonalization='cg'
> conv_thr = 1.0d-12
> mixing_beta = 0.7
> /
> ATOMIC_SPECIES
> C 12.0107 C.pbe-rrkjus.UPF
>
> CELL_PARAMETERS {angstrom}
> 2.5546 0.00000000 0.00000000
> 0.00000000 10.0 0.00000000
> 0.00000000 0.00000000 2.00000000
>
> ATOMIC_POSITIONS {angstrom}
> C 0.0 0.0 0.0
> C 1.2773 0.7375 0.000
> C 1.2773 2.2125 0.00
> C 0.0 2.95 0.0
>
> K_POINTS automatic
> 20 1 1 1 1 1
>
> ph.x
>
> phonons on kspace
> &inputph
> tr2_ph=1.0d-14,
> prefix='gnr_arm',
> ldisp=.true.,
> nq1=20, nq2=1, nq3=1
> amass(1)=12.0107,
> outdir='/home1/02420/pm24229/tmp/',
> fildyn='gnr_dyn',
> /
>
> Thanks,
> Prabhakar Marepalli
>
>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>
>
>
>
More information about the users
mailing list